Skip to content
Trending storyWatching

Coxon reads Anthropic chart as showing Mythos retrain needs far less compute

1 report1 sourceupdated 3 days ago

What happened

AI digest

The Decoder reported that Jacob Coxon, a former OpenAI and Anthropic researcher, said Anthropic can now retrain Mythos with a fraction of the compute previously required. He based this on a code pretraining compute multiplier chart from Anthropic's IPO roadshow, reading the curve as super-exponential growth in algorithmic efficiency. Anthropic notes the chart uses internal estimates and is illustrative only. Coxon also said algorithmic improvement is nearing automation and called for a pause on frontier model development next year so competitors can catch up.

Written by AI from the coverage · updated 3 days ago

Coverage

Follow the reports to see the story from different sides.

Nov 14
  1. The Decoder
    Anthropic can already retrain Mythos with "a fraction of the compute originally needed," says former researcher Coxon

    前 OpenAI 和 Anthropic 研究员 Jacob Coxon 称,Anthropic 已能用原先所需算力的一小部分重新训练 Mythos。其依据是 Anthropic IPO 路演中的代码预训练算力乘数图表,他将曲线解读为算法效率呈超指数增长;Anthropic 注明数据来自内部估算,仅供示意。Coxon 称算法改进接近自动化,并主张明年暂停前沿模型开发,让竞争者追赶。

Heat over time

Heat now 3·Comparable peak 10(Nov 14 21:00)·Comparable change over 24 hours -51%

02.557.510Nov 1421:00Nov 1512:00Nov 1604:00Nov 1619:00

The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.